Agilent 34401a AC current not reading correctly bellow 10 ma AC.

[jhormilla]

 I recently bought a Agilent 34401a from Ebay. I replaced all electrolytics and tantalums, all works great but the AC current bellow 10 ma AC, it was the same before replacing the caps and re-calibrating the 10 ma AC as per the calibration procedure. So at 10 ma AC and above it works normal and the readings match closely with a Keithley 2001 and a Siglent SDM3065X that I have, bellow 10 ma AC it rolls off, meaning 5 ma going in reads 2 ma, 1 ma going in reads 23 micro amps AC. AC volts, DC volts, DC amps and resistance all works fine. Any help would be appreciated.

Thank you.

[bdunham7]

That would appear to be entirely normal operation and within specification.  The AC ranges on the 34401A have less accuracy below 5% FS and are not specified below 1% FS.  In this case, 1% of 1A is 10mA, so you will have reduced accuracy below 50mA and no specified accuracy below 10mA.  The 34401A rolls the response to zero below 1% FS instead of just displaying residual counts.  It can be a bit misleading, but your meter is functioning normally in this regard.

If you want to measure small AC currents accurately you will need an external shunt.  I'm not sure why HPAK didn't use the internal DC shunts for AC. 

[floobydust]

This is normal for the 34401a, I'd stumbled onto it here : https://www.eevblog.com/forum/testgear/hp34401-ac-current-range-damaged-non-linear/

Also, there is a hidden 10mA AC range that can be enabled: https://www.eevblog.com/forum/testgear/agilent-34401a-hidden-10ma-ac-current-range/

[jhormilla]

Thank you both,
 I was wondering about the shunt used for AC ma not being the same used in the DC ma readings. I think its 5 ohms on DC ma and 0.1ohms on AC ma.

[TimFox]

In general, hardware accuracy for "true rms" measurements of AC waveforms suffers at low fractions of the full-scale.

[bdunham7]

Thank you both,
 I was wondering about the shunt used for AC ma not being the same used in the DC ma readings. I think its 5 ohms on DC ma and 0.1ohms on AC ma.

The default is that there is no AC ma range, just the 1A range which uses the 0R1 shunt.  The activation of the hidden range will apparently use the 5R shunt for a 10mAAC range, which should give pretty decent results up to 12mA.  I'll try it on mine as soon as I get a chance.

[bdunham7]

Also, there is a hidden 10mA AC range that can be enabled: https://www.eevblog.com/forum/testgear/agilent-34401a-hidden-10ma-ac-current-range/

Thanks for bringing that up!  I thought the 'hidden' range was a temporary digital amplification that could only be done remotely and was thus pretty worthless.  Now I see it is persistent and actually uses the 5R shunt, so I had to 'poke' mine and try it.  It works and seems reasonably accurate down to 1%FS or 100uAAC, although I would have to believe that the contrary to the Keysight note, the accuracy is not quite as good in equivalent terms as the 1AAC range.  Still, this gives the 34401A a reasonably accurate--although perhaps not quite up to 6.5 digit expectations--ACI measurement from 100uAAC to 3AAC.  The only quirk I see is that it will not autorange down into the 10mAAC range, but if you manually range down into it and reselect 'AUTO', it will autorange back up.  But not back down again.

For those of you without any GPIB capability, the poking can be done with a RS232-to-USB cable and HKJ's Testcontroller software.

[floobydust]

It seems to be a multimeter's worst performing function. I went nuts trying to measure around 1mA AC at 60Hz sine-wave, almost all of my multimeters could not do it. Either they have a dead zone and read near zero like the 34401a, or read way low like AN8008. Linearity is just not there.  A basic mains leakage-current test is required servicing some equipment.
It was hilarious, put every multimeter I had in series and in the end I could only trust my old '80's multimeters. Old Fluke 87-II did not do much better as I recall, it's got several digits of baloney as you wait for the true-RMS filter to settle, up to 30 seconds.

It's a real test that people do not do unfortunately.

[bdunham7]

It seems to be a multimeter's worst performing function. I went nuts trying to measure around 1mA AC at 60Hz sine-wave, almost all of my multimeters could not do it. Either they have a dead zone and read near zero like the 34401a, or read way low like AN8008. Linearity is just not there.  A basic mains leakage-current test is required servicing some equipment.
It was hilarious, put every multimeter I had in series and in the end I could only trust my old '80's multimeters. Old Fluke 87-II did not do much better as I recall, it's got several digits of baloney as you wait for the true-RMS filter to settle, up to 30 seconds.

It's a real test that people do not do unfortunately.

It is definitely an RTFM moment--even good meters can be surprisingly poorly specified for low ACI, and old bench meters especially so.  Although, the 34401A did manage to do almost as well as a Fluke 27 after the 'upgrade'.